Home AI News - AI Agents AI Adoption in Business: Why the Official Figures Don’t Match Up

AI Adoption in Business: Why the Official Figures Don’t Match Up

0
7
AI-generated illustration representing the adoption of artificial intelligence in small businesses

A single number is enough to divide the debate over artificial intelligence in business: depending on which organization publishes it, the AI adoption rate among France's small and midsize businesses (TPE-PME, meaning micro-businesses and small to midsize enterprises) varies by as much as three times over. France Num reports 40%, Banque de France cites 67%, Insee puts it at 18%, and Bpifrance comes in around 32%. The same phenomenon, four radically different snapshots.

The latest assessment of the Osez l'IA plan had already pointed to this difficulty in precisely measuring how AI is spreading through the economy. The new figures from France Num, published on September 28, 2026, raise the question again: should these statistics be trusted, and which one should we go by?

A Gap Explained by Methodology, Not by What's Happening on the Ground

According to the France Num barometer, 40% of French small and midsize businesses use at least one AI solution, up 14 points from a year earlier. The rate climbs to 53% among PME (10 to 249 employees) and drops to 30% among TPE (micro-businesses). The digital sector shows 72% adoption, specialized services 60%, compared with just 18% in agriculture.

These results, confirmed by several trade publications, don't really contradict the other surveys: they simply measure something different. Insee surveys a broader range of businesses and applies a stricter definition of “actual” AI use, which explains its lower rate of 18%. Banque de France, for its part, includes office tools that already come with built-in AI features (collaborative suites, messaging apps), which inflates its figure to 67%. Bpifrance Le Lab, finally, surveys a panel of business owners on a notion of “active and regular” use, close to 32%.

This spread isn't unique to France: it's the direct result of the lack of a standardized definition of what it means to “use AI” in a business, ranging from a one-off trial to a paid subscription to full integration into business processes.

What Real-World Usage Shows

Beyond the overall rate, the France Num barometer breaks down the actual uses among small and midsize businesses: 34% generate text, voice, or images with AI (compared with 22% in 2025 and 10% in 2024), 24% use chatbots or conversational assistants for their customers, 13% use AI for analyzing and sorting documents, and 11% use it to automate repetitive tasks. Only 19% of businesses pay for an AI solution (32% among PME): most usage therefore remains free or experimental.

This growth in concrete usage echoes what was already described in the article on still-limited productivity gains in 2026: adoption is progressing quickly, but the shift toward tools that are truly integrated into business processes—along the lines of the large-scale rollout of an AI agent at a major corporation—remains the exception rather than the rule at smaller companies.

Why These Numbers Matter for Small Business Owners

Beyond the statistical debate, these discrepancies have a practical consequence: they make any reliable sector-by-sector comparison difficult. A business owner who wants to know whether they're “behind” on AI compared with competitors first needs to check which definition and which scope lie behind the number they're looking at. When in doubt, it's better to look at detailed usage—content generation, customer service, task automation—than at the overall “adoption” percentage, which lumps together very different realities depending on which organization publishes it. These questions also touch on the economy of small and midsize businesses more broadly, and on how artificial intelligence technologies are spreading through France's productive fabric.

Frequently Asked Questions

Why do AI adoption rates in business vary so much depending on the source?

Because each organization (France Num, Insee, Banque de France, Bpifrance) uses a different scope of businesses and a different definition of “AI use”—ranging from a simple trial to a paid subscription to integration into an existing office tool.

What is the most recent AI adoption rate among French small and midsize businesses?

According to the France Num barometer from September 28, 2026, 40% of French small and midsize businesses use at least one AI solution, with a notable gap between PME (53%) and TPE (30%).

What are the most common AI uses in small businesses?

Generating text, voice, or images ranks first (34%), followed by chatbots for customer relations (24%), document analysis (13%), and automating repetitive tasks (11%).

Sources

Author Profile

Ethan Cole

This article was written with the help of artificial intelligence. Editorial policy

Previous articleMedical Franchise Fees and 15% Drug Reimbursement: What’s Changing for ALD Patients
Next articleCar Rental in Paris: The Practical Guide to Choosing Wisely

Editorial policy and use of artificial intelligence